What
Write a structured JSON log for each pipeline run — inputs, step results, token usage, approval decisions, timestamps.
Why
The audit trail is a core selling point of FixClaw, but currently logs only go to stdout. A JSON log file per run makes it easy to review, search, and archive pipeline history.
Scope
- Create a
logs/ directory (gitignored)
- After each pipeline run, write a JSON file:
logs/{pipeline-name}-{timestamp}.json
- Include: pipeline name, start/end time, each step's name/type/status/duration, token counts for AI steps, approval decisions and who approved
- Keep it append-friendly (one file per run, not one big file)
Helpful context
- Go's
encoding/json and os.WriteFile are all you need
- Look at where pipeline results are currently logged in
main.go
What
Write a structured JSON log for each pipeline run — inputs, step results, token usage, approval decisions, timestamps.
Why
The audit trail is a core selling point of FixClaw, but currently logs only go to stdout. A JSON log file per run makes it easy to review, search, and archive pipeline history.
Scope
logs/directory (gitignored)logs/{pipeline-name}-{timestamp}.jsonHelpful context
encoding/jsonandos.WriteFileare all you needmain.go